<d1b2>
<josHua> ngscopeclient takes quite some time to download 500Mpts out of this DHO4000, but, gosh, it is really nice to be able to scroll around an entire MIPI CSI2 frame from start to finish with no lag
Degi has quit [Ping timeout: 244 seconds]
Degi has joined #scopehal
<d1b2>
<azonenberg> 😄
<d1b2>
<azonenberg> we need a csi protocol decode, care to write one?
<d1b2>
<azonenberg> it's very similar to DSI so should be ~90% copy paste from the DSI filter
<d1b2>
<josHua> I would, if my probe were worth shit. even as slow as 150 MT/s, it looks like trash, and I only really am managing to eyeball where there are runs of zeroes and ones.
<azonenberg>
lol oh
<d1b2>
<josHua> I thought that littleprobe should be able to do better than this and I am currently very suspicious of the blu-tak that I have placed all over it
<azonenberg>
lol
<d1b2>
<josHua> yeah, ok, just looked at DSIPacketDecoder.cpp, that should be easy to write in terms of CSI2
<d1b2>
<azonenberg> yeah i just didn't have a CSI signal source to test against
<d1b2>
<azonenberg> note that the DSI decoder is several layers including the full framegrab to RGB scanlines
<d1b2>
<josHua> yeah, I would be happy to decode to packets. for almost all of my CSI debug, that is more useful to me
<d1b2>
<josHua> especially because debugging cameras, at best you can get bayer out, and getting reasonable looking images out of that is almost certainly out of scope for ngscopeclient
<d1b2>
<azonenberg> yeah with DSI the output is usually RGB or similar
<d1b2>
<azonenberg> but if i can get a scopesession with a working CSI decode and framegrab in it
<d1b2>
<azonenberg> i can attempt some higher level filters
<d1b2>
<josHua> best case scenario output is, like, a series of 12-bit-in-16-bit-container .raw files
<d1b2>
<azonenberg> certainly packets is the more immediately useful
<d1b2>
<josHua> then I can feed that to a mini-ISP that I'd hand-write or something
<d1b2>
<josHua> tomorrow if I manage to debug the actual problem I'm having and I don't have to worry about breaking the probe, I'll try to pick the blu-tak off of it and see if that helps the situation
<d1b2>
<josHua> then I should be able to get some nice CSI2 RAW12 captures that I can share
<d1b2>
<josHua> although hmm, I guess I have this board only instrumented with one probe, single-ended on a D-PHY data lane with a known UI time, so maybe that will not be as useful as one might hope
<d1b2>
<azonenberg> yeah i think the current decoder expects both halves of the d-phy data plus at least one side of the clock